The present invention relates to a method for reproducing pictures which reproduces video data of a moving picture coded with high efficiency by a MPEG (Moving Pictures Expert Group) method or the like in a jump reproducing mode.
The study for compressing a video signal of a moving picture with high efficiency to transmit, record and reproduce it has been actively carried out. For example, an attempt has been made to record, into a small disk, video data of a moving picture compressed with high efficiency. For the purpose of preparing an International Standard used in coding a video signal of a moving picture with high efficiency to compress video data, the MPEG successively proposes various kinds of data formats relating to video data compressed with high efficiency. The study and development have been also carried out for a practical system for transmitting, recording and reproducing video data compressed with high efficiency in accordance with the data formats proposed by the MPEG (a compressed video signal compressed with high efficiency by the MPEG method).
In a coding system (MPEG system) for video data of a moving picture using a recording medium for recording digital data such as CD-ROM, a predictive coding procedure is employed. There are three predictive methods, i.e., an intraframe predictive method which is applied to compress picture data to obtain a picture {I picture (Intra Pictures)} frame (hereinafter referred to I frame), an interframe predictive method for effecting an interframe prediction on the basis of picture data of a past frame which is applied to compress picture data to obtain a picture {P picture (Predicted Pictures)} frame (hereinafter referred to as P frame), and an interframe predictive method for effecting an interframe prediction on the basis of both picture data of a past frame and picture data of a future frame which is applied to compress picture data to obtain a picture {B picture (Bi-directional Prediction Pictures)} frame (hereinafter referred to as B frame). A predetermined header is added, to obtain picture coding data, to digital data in the state where the frames according to the aforementioned three picture modes are disposed on the time base in a predetermined mode.
In the MPEG method, the relationship between the compression rate of the picture data in the I frame, the compression rate of the picture data in the P frame and the compression rate of the picture data in the B frame is expressed as (the compression rate of the picture data in the I frame)&lt;(the compression rate of the picture data in the P frame)&lt;(the compression rate of the picture data in the B frame). It is necessary that in reproduction, the reproduction is carried out from a sequence header of an entry point, and that in reproduction of picture data of the B frame in which the prediction is carried out using the picture data of the past frame and the picture data of the future frame, the picture data of the future P frame which was used for the prediction of the picture data of the B frame is recorded before the B frame.
The coding according to the MPEG system as described above will now be described with reference to FIGS. 1A and 1B.
FIGS. 1A and 1B each show an example of arrangement of various frames in the MPEG method.
FIG. 1A shows the order of frames of picture before being coded, and FIG. 1B shows the order of frames of pictures after being coded.
A group of pictures (hereinafter referred to as GOP) indicates that a plurality of frames are arranged in a single form in order to facilitate the signal processing of picture data. In the FIG. 1 example, one GOP is composed of 15 frames, and at least one I frame is set every GOP.
In FIG. 1, I in frame Nos. 7 and 22 indicates I frame.
And P in frame Nos. 1 and 4 indicates P frame.
Further, in FIG. 1, B in frame Nos. 5 and 6indicates B frame.
In the coded case, a quantity of codes of the P frame and the B frame in which only a difference of the signal is coded is considerably small as compared with a quantity of codes of the I frame.
FIG. 1A shows the order of frames before being coded. In the coding, one GOP is composed, for example, of frames from frame Nos. 5 to 19.
FIG. 1B shows the order in which coded data are recorded in a recording medium. One GOP is composed of frames from frame Nos. 7 to 18, and the first frame constitutes I frame. The order recorded in the recording medium is the same as the order of frames displayed on a display unit after decoding but is slightly different from the order of a coding. Since decoding of the B frame requires data of the I frame and P frame, the P frame behind the B frame is recorded prior to the B frame at the time of coding, and is first read at the time of reproduction.
The reproduction of video from the video data of a moving picture coded with high efficiency by the MPEG method is carried out by various reproduction modes such as a slow reproduction mode, a static picture reproduction mode, a frame feed reproduction mode, a jump reproduction mode (a scan reproduction mode) and the like in addition to a normal reproduction mode.
Incidentally, in the case where the video data of a moving picture coded with high efficiency by the MPEG method is reproduced in the jump reproduction mode to obtain a reproduced video in a successively thinned-out state from the video data of a moving picture coded with high efficiency by the MPEG method, only the I frame having a sequence header and a GOP header disposed immediately therebefore is successively reproduced.
Conventionally, in the case where a reproduced video in a successively thinned-out state is obtained from video data of a moving picture coded with high efficiency by the MPEG method, the following operation is repeated so as to decode successive I frames. Such an operation comprises (1) a bit stream is continuously supplied to an MPEG video decoder; (2) in the MPEG video decoder, a sequence header is detected out of the bit stream supplied as described above, and video data of the subsequent I frame is decoded; (3) when the MPEG video decoder completes the decoding operation for the video data of one I frame, a central arithmetic processing unit receives a notice of the above completion; and whereby the central arithmetic processing unit searches a location where a next sequence header or a sequence header distant therefrom by a predetermined distance seems to be present to begin taking the bit stream therefrom, and the step returns to the above (1).
Incidentally, in order that a video reproduction system reproduces video data of a moving picture in a jump reproduction mode, when an operator gives an operating section a command so that the video reproduction system performs the operation in a jump reproduction mode, a disk drive device provided with an optical head causes the optical head to effect the seek operation in accordance with the command from the central arithmetic processing unit, data read by the optical head is once stored into a buffer memory, and the stored data is read from the buffer memory to continuously supply the bit stream to the MPEG video decoder. The MPEG video decoder sees the whole bit stream supplied thereto to detect the I frame therefrom and decode the video data. When the MPEG video decoder completes the decoding operation for the video data of one I frame, the central arithmetic processing unit receives a notice of such completion. The central arithmetic processing unit in turn issues a command to the disk drive device so as to cause the optical head to seek a location where an I frame which is next to the I frame already decoded or distant therefrom by a predetermined number of frames, seems to be present. However, in the case where the reproducing operation as previously mentioned is carried out to carry out the jump reproduction, since the I frame is detected by the MPEG video decoder, the video data of the I frame is decoded and after the completion of the decoding operation, a next seek operation is carried out, much time is taken other than the time such as time required for the seek operation of the optical head and time required till a servo operation enters a normal operation.
When video data of a moving picture coded with high efficiency by the MPEG method is reproduced in a jump reproduction mode, only the I frames with the video data compressed by applying the intraframe predictive method are successively reproduced from the video data of a moving picture coded with high efficiency by the MPEG method. Therefore, for example, if the central arithmetic processing unit sees all the bit streams so as to effect the detection operation of the sequence header, the detection of the sequence header can be possibly made in a short period of time. However, since the central arithmetic processing unit operates for processing many kinds of signals, enormous much time is taken in the case where the central arithmetic processing unit sees all the bit streams to effect the detection operation of the sequence header. For this reason, the detection operation of the sequence header has been heretofore carried out by the MPEG video decoder.
As described above, much waste time is required for the above-described detection operation of the sequence header, and as a result, much time is taken till the pictures of successive I frames are obtained. Thereby the number of pictures per unit time at the time of the jump reproduction decreases, resulting in an insufficient smoothness of the movement of a reproduced video. Measures for an improvement therefor has been demanded.
Further, in the case where a video reproduction system reproduces a picture by quick traverse/reverse from video data of a moving picture coded with high efficiency by an MPEG system, when an operator issues a command to an operating section so as to cause the video reproduction system to be operated in a reproduction mode of quick traverse/reverse, a central arithmetic processing unit in the video reproduction system sequentially controls the operation of various constituent portions of the video reproduction system in accordance with the steps shown in FIG. 2 illustrating the procedure for reproducing a picture by quick traverse/reverse from the video data of a moving picture coded with high efficiency by the MPEG method, thereby selecting I frame with a sequence header and a GOP header positioned immediately therebefore so as to obtain a quick traverse/reverse video.
More specifically, in the case where the reproducing operation of a picture by quick traverse/reverse video is carried out from video data of a moving picture coded with high efficiency by the MPEG method by the operation of constituent portions in the video reproduction system in accordance with the procedure as shown in FIG. 2, the following operation is carried out. A disk drive device provided with an optical head causes the optical head to effect the seek operation in accordance with a command of the central arithmetic processing unit, the data read by the optical head is once stored into a buffer memory, and the thus stored data is read from the buffer memory to continuously supply bit streams to an MPEG video decoder. The MPEG video decoder sees the whole bit stream supplied thereto to detect I frame therefrom and decode the video data of the I frame. When the MPEG video decoder completes the decoding operation in connection with the video data of one I frame, the central arithmetic processing unit receives a notice of such completion. The central arithmetic processing unit issues a command to the disk drive device so as to cause the optical head to seek a location where a next I frame seems to be present.
In the case where a picture is reproduced by quick traverse/reverse from the video data of a moving picture coded with high efficiency by the MPEG method as described above, the total time of (1) time required to skip to position of bit stream to start next reproduction, (2) time required to detect I frame after the start of reproduction of bit stream from the skipped position, and (3) time required to reproduce the entirety of the detected I frame is required every reproduction of the video data of successive I frames. However, as previously mentioned, the spacing between the successive I frames in the bit stream where I frame, P frame and B frame comprising the video data of a moving picture coded with high efficiency by the MPEG method are present in a mixed form is not constant but irregular. Therefore, the position at which skipping is effected as a position of bit stream to start next reproduction is, even at a fixed transfer rate, that skipping is made at a preceding position having enough room with respect to a position where I frame seems to be present so as to detect the I frame after the start of the reproduction of the bit stream from the skipped position. For this reason, the time required to detect the I frame as described above becomes extended, and as a result, the number of reproduced pictures per unit time decreases, resulting in an insufficient smoothness of the movement of the reproduced pictures. Measures for an improvement capable of making the aforesaid time as short as possible has been demanded.
Normally, in the case where video data is reproduced, by the reproduction system, from a recording medium having I, P and B frames recorded therein, the P and B frames cannot be decoded independently. For example, a picture of 30 frame/sec is decoded in unit of approximately 0.5 sec corresponding to 1 GOP, which is a minimum unit at which a complete picture can be restored.
In the case where the video data of a moving picture compressed as described above is reproduced, special reproductions such as a high speed reproduction for effecting reproduction at 2-fold speed, 3-fold speed, etc., and a reverse high speed reproduction for effecting high speed reproduction in a reverse direction in time are carried out. In this case, only the aforesaid independent frame (I frame) in which a reproduced picture is obtained only from codes within the frame is displayed on a display unit, and P frame and B frame are neither decoded nor displayed even if they are read.
In the conventional recording medium and the reproduction method therefor, the special reproduction is effected as follows. In case of a fixed transfer rate at which a quantity of codes delivered per unit time in the coding is substantially constant, a quantity of data corresponding to a length of 1 GOP capable of being independently reproduced or the number of sectors are calculated. A target point is accessed at an approximate skip distance, the whole data of the GOP is restored and only the I frame is displayed on the display unit.
Actually, however, since even the fixed transfer rate, the momentary data quantity is varied, it is necessary to start a reading considerably before the field to be accessed so as not to pass data which should be naturally accessed. Much time was taken for the access.
This poses problems that the complicated special reproduction cannot be accomplished, the special reproduction is limited to a random access, a search and the like, and the high speed reproduction such as the foldspeed reproduction cannot be accomplished.
Further, in case of a variable transfer rate coding in which a quantity of codes delivered per unit time in the coding is varied (for details, see Japanese Patent Application No. 339720/1992 filed by the present applicant, for example), the quantity of data per unit time considerably varies, and the aforementioned problems become further significant. In the conventional method for predicting a next access point from the quantity of data, the special reproduction was impossible.
As a method for solving the aforementioned problems, it is conceived that a Table is prepared which shows independent (I) frame access data indicative of an address of the first sector within a recording field in which independent frames are recorded with respect to all the independent frames, the data are recorded in a predetermined field of the recording medium, the data of the Table is first stored into a semiconductor memory such as a read S-RAM, etc. at the time of reproduction, and the special reproduction is carried out on the basis of the data of the Table. However, this poses a problem that a microcomputer in a reproduction system has to carry a huge burden of signal processing.
Generally, each sector of the recording medium is composed of subcode data in which control data such as a synchronous signal and an address, and a data field in which video data and the like are recorded.
As a further method for solving the aforementioned problems, there is conceived a method for recording the independent frame access data into a subcode field of a sector of a recording medium similarly to a music compact disk. In this case, address data recorded in one sector is not address data for all the independent frames but address data for several independent frames in the vicinity of the present frame is recorded.
However, in the above method, an interleave is applied for lowering the error rate between data of the subcode field and data of the data field. Therefore, the data of the subcode field is obtained only after the data of the data field is read, resulting in an increase in access time. This method is not suitable for the special reproduction.
For example, in CD-I (FMV) to which a musical compact disk is applied, not only compressed video and audio data but also various data are alternately recorded in unit of sector. It is impossible to obtain the access data for accessing pictures or sounds to be restored from only a specific sector. Reproduction order data is added every compressed data in predetermined unit to continuously restore the data.
As described above, the sector to be accessed next is indicated, and the special reproduction to some extent can be made. However, the data for access is recorded in the data field which requires complicate processing such as correction of error, deinterleave, etc. for the read data and so reading involves a variety of restrictions. The special reproduction is thus limited to the random access, search and the like, and the foldspeed reproduction or the like was actually impossible.